Otherwise we shift entire SF schedule for nothing...
Test: Open IME, inspect systrace
Bug: 118118435
Change-Id: I1352860a8f04271cd5dc72366ffc317175db5e3a
We handle changes to the leashes from the UI thread, but use
the same SurfaceControl wrapper object from the RenderThread with
SyncRtSurfaceTransactionApplier. This means that at the time
we release a SurfaceControl from the UI thread we might have already
scheduled a SyncRtSurfaceTransactionApplier to use it, and actually
that could be in the process of running, leading to racy access and
crashes. To fix this we release the SurfaceControl from the RenderThread
so that it happens behind all existing operations.
Bug: 151086678
Test: Existing tests pass.
Change-Id: I2308d1c64f3f368c32587f99ddfb9e05955f821f
In cases where we are replacing the client surface transparently (e.g.
preserved surfaces), we call reparentChildren to move client added
Surfaces of the preserved Surface to the new Layer. However we are calling
this on the WSA layer, and so we end up reparenting the BLAST Surface
to the new layer. But since we always construct a new BLAST surface on the
WM side, we end up with 2! Rather than try and juggle about when we need
to or need not to construct a BLAST surface and reparenting them across
preservedSurfaces, it seemed either to just use the BLAST surface as the
root of the reparentChildren operation.
Bug: 150013915
Test: Enable BLAST, use split-screen
Change-Id: I1357200b41d183c2331b684ff636dd40a3b98168
With Hierarchical animation, the animation layer will no longer
be a fixed fullscreen layer but animate on the parent container's surface.
In order to run a remote animation, the animation controller needs to
know bounds of the target relative to both its parent and the screen.
The CL includes:
1) RemoteAnimationTarget changes:
- Add localBounds field for indicating the target bounds which
the coodiates relatives to its parent.
- Add screenScreenBounds field to replace souceContainerBounds
to reflect the target bounds relatives to the screen.
- Mark position & sourceContainerBounds as deprecated.
2) Modified related places to set correct localBounds information.
Test: build / run, make sure installing the old version of launcher on
the this new platform change still compatible without crash.
Test: manual as follow steps:
- Launching app from launcher to split-screen secondary stack
- Swipe up to overview screen and drag TaskView to see if the TaskView
surface is shfted, expected is not.
Bug: 148780840
Change-Id: Id9dbf6de193ab73fe94bc24ef6a27edc93380a14
This adds some more config controls to containerTransaction so that
sysui divider can "freeze" the configuration w/h while it is
offset for IME.
Bug: 133381284
Bug: 149952263
Bug: 150400846
Test: Just after boot, open split with settings in secondary and
open the search activity. This should show IME and not
randomly hide it when background apps are started.
Also added a wmtest.
Change-Id: I731378f625079f8a1413469f71455650d9e59a50
SurfaceControl resources may be heavy weight, and so we explicitly
release them when finished. Specifically, we need to fix this to
keep the CloseGuard from complaining in strict-mode.
Bug: 150805473
Test: Existing tests pass
Change-Id: Ie3c609f79a953374a4d336465676cdc312679f93
In previous CL[1], WindowMetrics#WindowInsets is provided
with legacy insets, which are stable insets and system
window insets. With the new insets mode enabled, we
should provide insets with new insets API.
[1]: f65e022caa
Bug: 150095967
Test: atest WindowMetricsTests
Change-Id: I251313b4b8c033dbad8b54e58cd166e2e1665d37
1. Remove the internal using argument, ACTION_ARGUMENT_IME_ACTION_ID_INT.
2. Add more descriptions about this accessibility action, ACTION_IME_ENTER
Bug: 150831470
Test: atest AccessibilityTextActionTest
Change-Id: Icfebdf280e3441bf7b2b6f55c3febbe3a1d98de6
- Copy surface params builder to compat class
- Add calls to set background blur
- Make recents/app transition leashes effect layers so blur can be set
on them
Bug: 149792636
Test: Build with launcher with blurs enabled
Change-Id: I4cebcab090719c6a17f197a3cd4450d68e55b424
Forcibly-shown system bars are usually transient. Prevent them from
causing insets can reduce the UI shakiness, and can be compatible with
the behavior in the legacy insets mode.
Fix: 150582388
Test: atest InsetsSourceProviderTest InsetsStateControllerTest
InsetsPolicyTest
Change-Id: I3c0fa4fb7555b2f63e1c4849db7b169489ab64e4
Both showSoftInputUnchecked() and closeCurrenInput() rely on rootView to
obtain window token. If view root is null, window has already gone away
and IME control had been revoked. Trying to show or hide at this time
would be a no-op.
Bug: 149900693
Test: Manually using the steps mentioned in bug and verify that there is
no NPE.
Change-Id: I294bb2ec5395d7502a855bafbac672af069e9b4a
Merged-In: I294bb2ec5395d7502a855bafbac672af069e9b4a
(cherry picked from commit ba9b716a70)
Provide a recording insets controller before the window gets
created, and replay the commands once a view gets attached. This
allows the client to use the controller in Activity.onCreate.
Test: WindowInsetsControllerTests
Bug: 118118435
Change-Id: I1a825ecc4367c02b27f2d08cd5442325315d4f89
* autofill will cache the inline suggestions response until it receives
a start input view event from IME
* the data flow from IMS point of view is:
IMS#startViews and IMS#doStartInput (before calling onStartInputView)
->
[async] InlineSuggestionsRequestCallback#onInputMethodStartInputView()
--- process boundary ---
->
IMMS.InlineSuggestionsRequestCallbackDecorator
#onInputMethodStartInputView()
->
InlineSuggestionSession.InlineSuggestionsRequestCallbackImpl
#onInputMethodStartInputView()
* similar data flow for IMS#finishViews()
* this CL should not block IME's UI thread because it's only issuing a
new async IPC from IMS start/finish input view call that's running on
the UI thread.
* there should not be performance impact on IMEs if autofill inline
integration is not active
Test: manual verification, atest EditorInfoTest
Test: atest android.autofillservice.cts.inline, with two failing cases:
InlineAugmentedLoginActivityTest#testAugmentedAutoFill_twoDatasetThenFilledSecond
and InlineAugmentedLoginActivityTest#testAugmentedAutoFill_oneDatasetThenFilled
due to the test itself being broken, I'll fix the test in a separate patch
Bug: 149522488
Bug: 149442582
Change-Id: I2faa3577b9f95a122f26a6d7fa7822a769a51e34
